Loading page...Загружаем страницу...正在加载页面...
AtlasRepo

The Inbox Autopilot: What 60 Days of Letting an n8n Workflow Triage My Email Taught Me

说明与用途

使用 n8n 和 LLM 自动化 Gmail 邮件 triage 与草稿生成,每天节省 35 分钟,采用置信门和语气模板。

问题

Daily email triage consumed 45 minutes each morning, requiring manual decisions on urgency, reply necessity, and noise.

结果

The workflow reduced triage time by ~35 minutes daily, introduced a confidence gate to handle uncertainty, and used per-bucket tone templates to improve draft appropriateness. Over 60 days, no incidents occurred.

工作流

  • Trigger on new Gmail messages
  • Strip signatures, disclaimers, and quoted history via Code node
  • Classify email into buckets (urgent, needs-reply, waiting-on-me, newsletter, noise) using LLM
  • Apply confidence gate; emails below 0.75 go to unsure bucket
  • For needs-reply, generate draft reply in author’s voice via second LLM call
  • Store drafts in Gmail drafts folder
  • Send urgent pings via Slack/Telegram
  • Auto-archive newsletters
  • Morning review: search Gmail drafts newer_than:1d, approve or edit

证据

Every morning I did the same thing: open Gmail, scan 60–80 new emails, decide what needed a reply, what could wait, and what was noise. It took 45 minutes on a good day.

来源

It now saves me about 35 minutes every morning.

来源

The workflow never sends email. Draft-only.

来源

The confidence gate (the part I'd copy anywhere) ... Below 0.75, the email goes into a sixth bucket: unsure – always shown to the human, never auto-filed.

来源

  • 记录日期: 2026-10-16T10:40:59.613Z